Skip to content

knowledge: consolidate 30 open flush PRs (#47–#110) into one reconciled state — 53 new pages, 51 amended - #112

Open
choiyounggi wants to merge 90 commits into
mainfrom
knowledge/consolidation-20260818
Open

knowledge: consolidate 30 open flush PRs (#47–#110) into one reconciled state — 53 new pages, 51 amended#112
choiyounggi wants to merge 90 commits into
mainfrom
knowledge/consolidation-20260818

Conversation

@choiyounggi

Copy link
Copy Markdown
Owner

What

The 30 open knowledge/* PRs (2026-08-06 → 2026-08-17) each rewrote log.md/.dev-loop/INGEST_REPORT.md and several amended the same wiki pages, so per-PR merge meant a conflict cascade at every step. Following the established #17#40 consolidation protocol, this branch carries the reconciled end-state; the 30 source PRs will be closed in its favor. Full report: .dev-loop/INGEST_REPORT.md.

Verified best-practice

Every adopted page's sources were carried from its originating PR's flush, where they were live-verified at flush time; no new URLs were introduced during consolidation — every added page arrived whole from exactly one branch, and every amended-page hunk came from a branch-side merge hunk, not hand-written text. Confidence fields kept as the originating flushes set them. wiki-lint-prohibitions.js: 71 directives / 0 violations; 3 vague-qualifier hits in new-page directive sentences rewritten as stated conditions (quoted source text untouched).

Existing-layer check

Pages read: testing-quality-tests-that-cannot-fail, testing-quality-behavior-not-implementation, testing-quality-guard-shape-vs-consequence, testing-quality-harness-reverse-controls, testing-quality-policy-at-several-return-sites, testing-data-test-data-and-isolation, backend-common-change-impact-call-site-enumeration, backend-common-change-impact-inserting-a-guard-before-an-existing-side-effect, backend-common-change-impact-cross-module-consumer-census, backend-java-jpa-raw-jdbc-inside-a-jpa-transaction, infrastructure-agent-orchestration-worktree-isolated-workers, infrastructure-agent-orchestration-shared-run-state, infrastructure-agent-orchestration-session-completion-gates, qa-document-verification-spec-document-gates, qa-document-verification-editing-a-gated-document, qa-exploratory-guard-true-path-coverage, qa-environments-headless-browser-bot-blocking, qa-process-defect-class-resweep-after-review, qa-process-scope-purity-checks (conflict-resolution and qualifier-polish reads on the final tree)

Open-PR check

This PR is the reconciliation of the open-PR backlog: all 30 open knowledge/* PRs (#47 #49 #50 #51 #52 #55 #56 #57 #58 #61 #62 #64 #66 #68 #69 #72 #73 #74 #76 #78 #79 #80 #86 #91 #92 #95 #101 #103 #104 #110) are merged into this branch and will be closed in its favor once it lands. No other open knowledge PRs remain; #111 (WebMCP wiki pages) was merged to main first and this branch is based on it.

Routing decision

  • 53 new pages, 51 amended pages, 12 indexes updated.
  • New categories: backend/common/ml (mape-aligned-point-prediction), backend/python/packaging (data-files-and-install-paths), security/incident-response (verifying-assumed-security-agents, process-identity-by-path-and-hash). All other new pages route into existing categories; INDEX.md domain summaries updated for backend, frontend, security.
  • Verification: bats 563 ok / 0 failed.

🤖 Generated with Claude Code

https://claude.ai/code/session_01Rb8k418PNt7LEbrtXJX6pJ

dch0202 and others added 30 commits August 6, 2026 13:04
…downgrade 2 confidences to field-tested, correct 4 mechanism claims
Drains 8 queued candidates: 2 new pages (qa/deliverables/quantitative-claims-in-a-published-document,
security/data/commit-identity-in-public-repos), 3 merges into existing pages
(debugging hypothesis-testing, agent-orchestration worktree-isolated-workers and
pane-delivery-confirmation), 3 drops (2 already merged on main verbatim, 1 out of
wiki scope). One candidate's stated mechanism was corrected by reproduction against
guardrails 1.2.0 before ingest rather than overwriting the conflicting page.
…ing assertions, query state vs fetch state, python text-io encoding)
…, regex-reach claim, equivalence gate, missing state cells)
…ntional warnings); 2 dropped as in-flight duplicates
New: infrastructure/agent-orchestration/unattended-worker-questions — an
unattended worker's in-band question reaches nobody: with a TTY the chooser
waits indefinitely while liveness checks pass, without one it self-answers
empty. Classify a live-terminal stall from the terminal tail before restarting,
unblock with an allowlisted key sequence, re-send the prompt the chooser
swallowed, and prevent recurrence with a durable out-of-band question record.

Dropped: the worktree_escape read-only-escalation candidate (in flight in #47
and #51) and the Orca dispatch-binding taxonomy candidate (in flight in #51/#47).
log.md records a contradiction between #47 and #51 on the guardrail mechanism,
settled against the rule source in favour of #51.
…ring-assertions (step 2 + edges); tokenizer over regex, measured
…gs) to both new pages

- suppression: the prescribed shell shape was wrong. `notify "$@" || return 1` at
  a script's top level errors and falls through to the marker write, exiting 0 —
  reproducing the defect the page exists to prevent. Prescribe the shape the cited
  repo actually ships (`if notify "$@"; then <write mark>; fi`) and table the
  three shell shapes including set -e propagation.
- suppression: the notify package docs state the stages' contracts, not their
  order; present the ordering as following from SetNotifiesStage's precondition.
- suppression: drop the 'not to a prior attempt' clause the current
  repeat_interval docs no longer support.
- suppression/consumer: fold each index 'load when' promise into 'When this applies'.
- consumer: add the non-deterministic-consumer edge (the two-run differ assertion
  passes vacuously); pin the loud/silent read sites by file:line; inline the
  reproduction's formula; convert the one prohibition outside 'Instead of'.
- ts.createSourceFile yields comments === undefined and no comment nodes, so
  naming it as the stripper is a silent no-op; name getLeadingCommentRanges,
  espree {comment:true} and @babel/parser parse().comments instead (measured).
- step 6's 'green by construction' covered only the comment half of the control;
  a reformat still reddens a bounded pattern with no comment involved (measured).
- the no-parser fallback must cut each line to EOL — filtering whole comment
  lines leaves trailing comments, the shape that actually leaks (measured).
- N must be measured on the stripped subject, not the raw source.
- label the fixture measurement by the removal that was actually run.
- testing/mocking/captured-call-arguments: record the whole call, assert every
  caller-decided argument; split the constant's value from the call site passing
  it on; re-prove per argument after extracting a resolver.
- databases/indexing/trigram-index-short-patterns: a pg_trgm wildcard segment
  under 3 chars yields no extractable trigrams, so GIN degenerates to a
  full-index scan while the plan still reads Bitmap Index Scan.
- backend/java/jpa/raw-jdbc-inside-a-jpa-transaction: @transactional(timeout)
  reaches JdbcTemplate only via a bound ConnectionHolder, and the JPA and raw
  paths raise different Spring exceptions (boundary at Spring Framework 7.0.0).
…under test

The queue held 4 pending rows, not 3: wc -l undercounted a file whose last line
had no trailing newline. testing/quality/default-values-under-test — a spec-named
constructor default is guarded by neither a mechanism test that passes the value
in nor a defaults-constructing test that never exercises it; push it to its
observable point and require red in both mutation directions (the grow direction
is the one no incidental test catches).
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants